Skip to content

Commit 5ec932b

Browse files
format with black
1 parent ed23c10 commit 5ec932b

File tree

1 file changed

+15
-28
lines changed

1 file changed

+15
-28
lines changed

src/unstructured_client/_hooks/custom/split_pdf_hook.py

Lines changed: 15 additions & 28 deletions
Original file line numberDiff line numberDiff line change
@@ -7,25 +7,25 @@
77
import logging
88
import math
99
import os
10-
from concurrent.futures import ThreadPoolExecutor, Future
11-
from typing import Optional, Tuple, Union, Generator
10+
from concurrent.futures import Future, ThreadPoolExecutor
11+
from typing import Generator, Optional, Tuple, Union
1212

1313
import requests
14+
from pypdf import PdfReader, PdfWriter
15+
from pypdf.errors import PdfReadError
1416
from requests.structures import CaseInsensitiveDict
1517
from requests_toolbelt.multipart.decoder import MultipartDecoder
1618
from requests_toolbelt.multipart.encoder import MultipartEncoder
17-
from pypdf import PdfReader, PdfWriter
18-
from pypdf.errors import PdfReadError
1919

2020
from unstructured_client._hooks.custom.common import UNSTRUCTURED_CLIENT_LOGGER_NAME
2121
from unstructured_client._hooks.types import (
22-
BeforeRequestContext,
23-
AfterSuccessContext,
2422
AfterErrorContext,
25-
SDKInitHook,
26-
BeforeRequestHook,
27-
AfterSuccessHook,
2823
AfterErrorHook,
24+
AfterSuccessContext,
25+
AfterSuccessHook,
26+
BeforeRequestContext,
27+
BeforeRequestHook,
28+
SDKInitHook,
2929
)
3030
from unstructured_client.models import shared
3131

@@ -61,9 +61,7 @@ def __init__(self) -> None:
6161
self.partition_responses: dict[str, list[requests.Response]] = {}
6262
self.partition_requests: dict[str, list[Future[requests.Response]]] = {}
6363

64-
def sdk_init(
65-
self, base_url: str, client: requests.Session
66-
) -> Tuple[str, requests.Session]:
64+
def sdk_init(self, base_url: str, client: requests.Session) -> Tuple[str, requests.Session]:
6765
"""Initializes Split PDF Hook.
6866
6967
Args:
@@ -314,9 +312,7 @@ def _parse_form_data(self, decoded_data: MultipartDecoder) -> FormData:
314312
for part in decoded_data.parts:
315313
content_disposition = part.headers.get(b"Content-Disposition")
316314
if content_disposition is None:
317-
raise RuntimeError(
318-
"Content-Disposition header not found. Can't split pdf file."
319-
)
315+
raise RuntimeError("Content-Disposition header not found. Can't split pdf file.")
320316
part_params = self._decode_content_disposition(content_disposition)
321317
name = part_params.get("name")
322318

@@ -327,9 +323,7 @@ def _parse_form_data(self, decoded_data: MultipartDecoder) -> FormData:
327323
filename = part_params.get("filename")
328324
if filename is None or not filename.strip():
329325
raise ValueError("Filename can't be an empty string.")
330-
form_data[PARTITION_FORM_FILES_KEY] = shared.Files(
331-
part.content, filename
332-
)
326+
form_data[PARTITION_FORM_FILES_KEY] = shared.Files(part.content, filename)
333327
else:
334328
form_data[name] = part.content.decode()
335329

@@ -377,9 +371,7 @@ def _call_api(
377371
raise RuntimeError("HTTP client not accessible!")
378372
page_content, page_number = page
379373

380-
new_request = self._create_request(
381-
request, form_data, page_content, filename, page_number
382-
)
374+
new_request = self._create_request(request, form_data, page_content, filename, page_number)
383375
prepared_request = self.client.prepare_request(new_request)
384376

385377
try:
@@ -469,9 +461,7 @@ def _prepare_request_payload(self, form_data: FormData) -> FormData:
469461
payload.update(updated_parameters)
470462
return payload
471463

472-
def _create_response(
473-
self, response: requests.Response, elements: list
474-
) -> requests.Response:
464+
def _create_response(self, response: requests.Response, elements: list) -> requests.Response:
475465
"""
476466
Creates a modified response object with updated content.
477467
@@ -490,9 +480,7 @@ def _create_response(
490480
setattr(response_copy, "_content", content)
491481
return response_copy
492482

493-
def _await_elements(
494-
self, operation_id: str, response: requests.Response
495-
) -> Optional[list]:
483+
def _await_elements(self, operation_id: str, response: requests.Response) -> Optional[list]:
496484
"""
497485
Waits for the partition requests to complete and returns the flattened
498486
elements.
@@ -525,7 +513,6 @@ def _await_elements(
525513
flattened_elements = [element for sublist in elements for element in sublist]
526514
return flattened_elements
527515

528-
529516
def _clear_operation(self, operation_id: str) -> None:
530517
"""
531518
Clears the operation data associated with the given operation ID.

0 commit comments

Comments
 (0)